The Centers for Medicare and Medicaid Services (CMS) and other large payers have launched numerous initiatives aimed at improving the quality of medical care, including public reporting, pay-for-performance, and payment reforms designed to increase coordination of care. Whether such initiatives will improve clinical outcomes and reduce costs is uncertain, however. In this context, the candidate (Dr. Lena M. Chen) seeks to develop a better measure of quality for inpatient medical care, and to understand the relationship between hospital quality and total costs of care. The candidate, a hospitalist and health services researcher at the University of Michigan, will leverage this proposal to develop her research agenda, and develop into an independent investigator. During the period of support, she will pursue didactic instruction in several disciplines, including doctoral-level courses in advanced statistical methods and economics. She will also have ample opportunity for mentored, project-based learning, including the hands-on application of advanced statistical modeling and econometric techniques. The proposed research plan has two aims: Aim 1. To develop and validate a composite quality measure for acute myocardial infarction, congestive heart failure, and pneumonia. Using national Medicare claims data, the candidate will use empirical Bayes techniques to develop a composite measure of quality. Model inputs will include structural variables, process measures, and risk- adjusted outcomes. Model outputs will be risk-adjusted, hospital-specific, condition-specific estimates of 30- day mortality. We will also pilot a parallel and supplemental combined outcome measure: death or the inability to live independently after hospital discharge. Aim 2. To understand the relationship between hospital quality and costs. The candidate will examine the condition-specific association between hospital quality and total Medicare payments, both with and without price adjustment. In addition to assessing overall payments around episodes of care, she is also interested in understanding the relationship between quality and specific types of costs, including those for index hospitalizations, readmissions, post-discharge payments, and ancillary care. She posits that high quality hospitals will have higher payments associated with greater post-discharge ancillary care, but lower payments overall, largely due to fewer readmissions. Findings from Aim 1 will inform efforts to improve hospital quality by combining multiple measures into a single metric better tied to clinical outcomes. Findings from Aim 2 will provide payers with insight into the impact of quality improvement efforts on costs.